martin-ueding
commented
7 months ago Yes, the GPX format supports multiple tracks with multiple segments. So OpenTracks exports each activity as one track? Are there names for activities included somewhere?
At the moment I just load all points from all segments from all tracks:
def read_gpx_activity(path: pathlib.Path, open) -> pd.DataFrame:
points = []
with open(path) as f:
gpx = gpxpy.parse(f)
for track in gpx.tracks:
for segment in track.segments:
for point in segment.points:
points.append((point.time, point.latitude, point.longitude))
return pd.DataFrame(points, columns=["time", "latitude", "longitude"])So that part should be fine, you end up with one huge activity. It wouldn't be what you want, but it should work.
It is a bit strange that it crashes. I'd need to take a look into the data types. If you want, you could just send me your GPX file to mu@martin-ueding.de and I'll take a look. If you don't want to send me your whole activity database, that's fine as well. I've made a small change to the code such that we get a bit more debug output. Just do a git pull and try again. Then there should be some additional output of the parsed data as a table and also the data types of the types. I would be interested what the time column is marked as.
pstorch
I tried to read .gpx files from OpenTracks. There is the possibility to export all tracks in one file. Then I get this error on startup:
I don't know if it is really the "multiple tracks" file which causes this error or if just one of the many tracks I exported is responsible for it.